MRI敷霜征鉴别乳腺良恶性小肿块的价值及其病理组织学分析

The Value of Blooming Sign on MRI in Distinguishing Malignancy from Benign Small Breast Masses and Its Radiologic-pathologic Correlation Analysis

【目的】探讨MRI敷霜征鉴别乳腺良恶性小肿块的价值,并分析其病理组织学改变。【方法】回顾性收集2016年6月~2020年9月我院行MRI检查并经病理证实的554例最大径线≤2 cm的乳腺小肿块患者资料(恶性291例,良性263例)。将患者的临床和MR特征进行单因素及多因素回归分析,筛选乳腺癌独立危险因素。基于独立危险因素构建两个诊断模型(模型1包含敷霜征,模型2不包含敷霜征)。采用ROC曲线评估两个模型的诊断效能。观察所有小肿块瘤周组织的病理学改变,分析敷霜征的病理学基础。【结果】恶性组199例(68.4%)、良性组25例(9.5%)出现敷霜征,两组间差异具有统计学意义(P<0.05)。单因素及多因素回归分析显示,年龄、病灶径线、边缘、ADC值、时间-信号强度曲线类型、敷霜征为乳腺癌的独立危险因素,OR估计值分别为1.065、4.515、2.811、0.013、3.487和13.894,OR 95%CI分别为(1.034,1.097)、(2.368,8.608)、(1.954,4.045)、(0.004,0.049)、(2.087,5.826)和(7.026,27.477)。模型1(包含敷霜征)对乳腺小肿块的鉴别诊断效能优于模型2(不包含敷霜征)(AUC:0.938 vs.0.897,P<0.05)。病理组织学分析显示敷霜征与瘤周炎性细胞浸润、血管增生相关。【结论】MRI敷霜征有助于乳腺良恶性小肿块的鉴别诊断,其病理组织学基础可能为瘤周炎性细胞浸润、血管增生。

【Objective】To determine the value of MRI blooming sign in differentiating benign and malignant small breast masses and investigate its radiologic-pathologic correlation.【Methods】This retrospective study included 554 small breast masses(291 malignant and 263 benign)which were≤2 cm and validated by pathology analysis between June 2016 and September 2020.All 554 patients underwent breast MRI.The clinical characteristics and MR features were analyzed.Univariate and multivariate regression analysis were performed to identify the independent risk factors of breast cancer.Two diagnostic models were constructed based on independent risk factors(model 1 included blooming sign and model 2 didn’t).ROC curve was used to evaluate the diagnostic performances of the two models.The histological changes of peritu⁃moral tissues in all small masses were analyzed.【Results】The blooming sign was positive in 199 cases(68.4%)of the ma⁃lignant masses and 25 cases(9.5%)of the benign ones(P<0.05).Univariate and multivariate regression analysis showed that age,lesion diameter,margin,ADC value,time signal intensity curve type and blooming sign were independent risk factors for breast cancer.Odds ratio were 1.065,4.515,2.811,0.013,3.487 and 13.894,respectively.Their correspond⁃ing 95%CI were(1.034,1.097),(2.368,8.608),(1.954,4.045),(0.004,0.049),(2.087,5.826)and(7.026,27.477),respectively.The diagnostic performance of model 1(blooming sign included)was better than that of model 2(blooming sign not included;AUC:0.938 vs 0.897,P<0.05).Histopathological analysis showed that the blooming sign was related to peritumoral lymphocyte infiltration and vascular proliferation.【Conclusions】MRI blooming sign is helpful for distinguishing breast cancer from benign masses.The correlated histopathological basis may be peritumoral lymphocyte in⁃filtration and neovascularization.